Hi,

We are a new customer to vMix and are experiencing an issue where vMix seems to lock up the resources to our network interface completely. Details are below:

PC Specifications:
-CPU: Intel i9-10850K @ 3.6 GHz
-RAM: 64 GB
-Graphics: Radeon RX 570 (x2)
-OS: Windows 10 Pro

Intranet Path Specifications:
1 GB Ethernet NIC (Gigabyte embedded NIC in motherboard for PC described above) <--> Cat 6A <--> Ubiquiti 24-port managed switch <--> 10GB/e fiber <--> Ubiquiti 48-port managed switch <--> 10GB/e fiber <--> Ubiquiti Dream Machine Pro (router) <--> Cat 6A <--> ISP (cable) modem <--> ISP service <--> RTMP destination

ISP Specifications:
-600 MB/s (down)
-35 MB/s (up)

Full error history w/ video:
https://drive.google.com/drive/folders/1Uz5ThYOCm7Qtz2FyG1HYILwXpFsI-B76?usp=sharing

Error details:
========================================
24.0.0.71 - FFMPEG1.Exited - 10/8/2021 4:58:22 PM
========================================
kzzzzzx: Stream1: Unable to connect to streaming server.
Check the internet connection or Stream URL and try again. ---> kzzzzzx: Streaming
frame= 602 fps= 31 q=27.0 size= 11067kB time=00:00:20.05 bitrate=4521.8kbits/s
av_interleaved_write_frame(): Unknown error

frame= 572 fps= 31 q=27.0 size= 10558kB time=00:00:19.02 bitrate=4546.0kbits/s
frame= 587 fps= 31 q=27.0 size= 10795kB time=00:00:19.53 bitrate=4526.1kbits/s
frame= 602 fps= 31 q=27.0 size= 11067kB time=00:00:20.05 bitrate=4521.8kbits/s
av_interleaved_write_frame(): Unknown error
Last message repeated 1 times
[flv @ 0b005200] Failed to update header with correct duration.
[flv @ 0b005200] Failed to update header with correct filesize.
frame= 605 fps= 31 q=27.0 Lsize= 11125kB time=00:00:20.13 bitrate=4526.0kbits/s
video:10809kB audio:299kB subtitle:0kB other streams:0kB global headers:0kB muxing overhead: 0.147163%
[libx264 @ 0b006160] frame I:11 Avg QP:22.30 size:139741
[libx264 @ 0b006160] frame P:594 Avg QP:24.50 size: 16861
[libx264 @ 0b006160] mb I I16..4: 79.4% 0.0% 20.6%
[libx264 @ 0b006160] mb P I16..4: 18.9% 0.0% 0.0% P16..4: 10.6% 1.5% 0.5% 0.0% 0.0% skip:68.4%
[libx264 @ 0b006160] coded y,uvDC,uvAC intra: 78.2% 10.3% 2.0% inter: 4.6% 4.0% 0.2%
[libx264 @ 0b006160] i16 v,h,dc,p: 24% 16% 55% 6%
[libx264 @ 0b006160] i4 v,h,dc,ddl,ddr,vr,hd,vl,hu: 28% 24% 19% 4% 5% 4% 6% 4% 6%
[libx264 @ 0b006160] i8c dc,h,v,p: 83% 9% 7% 0%
[libx264 @ 0b006160] Weighted P-Frames: Y:0.0% UV:0.0%
[libx264 @ 0b006160] kb/s:4578.28
Conversion failed!

--- End of inner exception stack trace ---

When this error occurs, the PC that vMix is installed on is completely unreachable from our internal network until after a full reboot of the PC. Other PC's however, have no issue connecting to our RTMP service provider or anywhere else for that matter. We can run other livestream software packages from this same hardware over the same network and to the same RTMP destination with no issues. We like the vMix GUI better than others, but will not be able to continue a subscription unless this issue is solved.

Also - there are no logs generated in Windows Event Manager when this happens - only after the fact when other services are unable to connect to anything.

Please let me know what other information you need to help us bring this to a resolution. If there are some settings on vMix that we have misconfigured, that's fine - we just need to know how to fix it.

Thanks!

-Jason

Is the NIC actually using a Gigabyte driver? I would update the driver from the OEM (Intel, Broadcom, etc.) because that's easy to do. Is the lockup intermittent or can you replicate on demand? This may be a symptom and not the cause. I'm wondering about the AMD cards too. Could be happening when encoding starts.

Thanks for your suggestion. The NIC driver is the latest Intel driver. I can replicate on demand and have obtained a PCAP during the crash to see if there's anything in there that would give a clue as to why this happens. Initially I thought that maybe excessive logical ports were being opened to manage the stream and then entered a TIME_WAIT or CLOSE_WAIT status that prevented reuse for a bit. That is not the case.

The only thing I have found so far is malformed RTMP requests sent from vMix on every PCAP of a crash that I have viewed. There are significantly more malformed packets sent on when vMix crashes than when it does not crash, but I can't seem to correlate that with adequate evidence to the locking up of the entire Ethernet NIC where no software on the machine can use that interface until a reboot. Something is seriously wrong here, despite the non-Nvidia graphics cards.

We're going to discontinue use of vMix for now - we lost another broadcast last night and I'm not willing to go down that road any further until a remedy is found. I do plan to switch to a single Nvidia graphics card but have serious concerns regarding that being a potential fix for these symptoms.

Thanks to everyone so far for your comments.

Yeah - I saw that... seems very strange to me, but maybe just because those configurations have not been tested by vMix, and therefore cannot be officially recommended? IDK... still seems strange.

Thanks!